KRAS G13D sensitivity to neurofibromin-mediatedGTP hydrolysisDana Rabaraa,1, Timothy H. Trana,1, Srisathiyanarayanan Dharmaiaha, Robert M. Stephensa, Frank McCormicka,b,2,Dhirendra K. Simanshua,2, and Matthew Holderfielda,2,3

aNCI RAS Initiative, Cancer Research Technology Program, Frederick National Laboratory for Cancer Research, Frederick, MD 21701; and bHelen DillerFamily Comprehensive Cancer Center, University of California, San Francisco, CA 94158

Contributed by Frank McCormick, September 16, 2019 (sent for review May 17, 2019; reviewed by Karen Cichowski and Kenneth D. Westover)

KRAS mutations occur in ∼35% of colorectal cancers and promotetumor growth by constitutively activating the mitogen-activatedprotein kinase (MAPK) pathway. KRASmutations at codons 12, 13,or 61 are thought to prevent GAP protein-stimulated GTP hydro-lysis and render KRAS-mutated colorectal cancers unresponsive toepidermal growth factor receptor (EGFR) inhibitors. We reporthere that KRAS G13-mutated cancer cells are frequently comu-tated with NF1 GAP but NF1 is rarely mutated in cancers withKRAS codon 12 or 61 mutations. Neurofibromin protein (encodedby the NF1 gene) hydrolyzes GTP directly in complex with KRASG13D, and KRAS G13D-mutated cells can respond to EGFR inhibi-tors in a neurofibromin-dependent manner. Structures of the wildtype and G13D mutant of KRAS in complex with neurofibromin(RasGAP domain) provide the structural basis for neurofibromin-mediated GTP hydrolysis. These results reveal that KRAS G13D isresponsive to neurofibromin-stimulated hydrolysis and suggestthat a subset of KRAS G13-mutated colorectal cancers that areneurofibromin-competent may respond to EGFR therapies.

KRAS | G13D | NF1 | EGFR | GTPase

The RAS family of protooncogenes cycle between active GTP-bound and inactive GDP-bound states in response to mito-genic stimuli. In the GTP-bound state, RAS proteins bind to andactivate the RAF/MAPK (mitogen-activated protein kinase) andPI3K pathways to promote cell-cycle progression. The rate ofintrinsic GTPase reaction of RAS is slow (1). Thus, the activeGTP-bound state of RAS proteins is primarily regulated byRasGAPs (Ras GTPase-activating proteins), which increase therate of GTP hydrolysis by ∼105-fold (2). Two well-characterizedRasGAPs are neurofibromin (the protein is referred to here asNF1, encoded by the NF1 gene) and RASA1 (also calledp120GAP) (3). RasGAPs achieve this by providing an arginineresidue (also known as an arginine finger) in the nucleotide-binding pocket of RAS, where it stabilizes and orients the cat-alytic residue, Q61, for an inline nucleophilic attack on thegamma-phosphate of GTP (4–6).Biophysical analysis of the GAP-mediated GTPase reaction in

the RAS–RasGAP complex has suggested 3 key steps for thereaction mechanism (7). In the initial step, RasGAP interactswith active GTP-bound RAS and forms a ground-state complex.During this step, the arginine finger remains exposed to the aqueousenvironment. The ground state is followed by the transition state,where the arginine finger positions itself in the active site, triggeringthe cleavage of GTP and formation of protein-bound Pi interme-diates. The last and rate-limiting step of the GAP-mediated GTPasereaction involves the release of Pi from the active site. So far, thestructural information on the RAS–RasGAP complex is limited tothe transition-state structure of the HRAS complexed with theGAP-related domain (GRD) of RASA1 (HRAS–RASA1GRD) inthe presence of GDP and AlF3, where AlF3 mimics the cleavedgamma-phosphate during the cleavage reaction (4).GAP-mediated GTP hydrolysis is frequently disrupted in hu-

man cancers by activating point mutations of RAS genes. KRASis the most frequently mutated of the 3 RAS isoforms. Mutations

are commonly observed near the nucleotide-binding pocket atglycine 12, glycine 13, or glutamine 61 (8). Codon 12 mutationspredominate across lung, pancreas, and colon, while codon 13mutations largely appear in colorectal cancers (CRCs). Allmutations in this region are thought to prevent formation of theRasGAP transition state by blocking the arginine finger fromaccessing the GTP terminal phosphate, thereby preventingRasGAP-mediated GTP hydrolysis. NF1 mutations are alsofrequent in malignant peripheral nerve sheath tumors (9) andin a fraction of lung and colorectal tumors but are thought tobe functionally redundant and mutually exclusive withactivating KRAS mutations.It is unknown why the KRAS G13D mutation appears almost

exclusively in gastrointestinal cancers and is rare in lung and

Significance

The protooncogene KRAS, a GTPase, is responsible for acti-vating the MAPK pathway. Cancer mutations at codons 12, 13,or 61 are thought to stabilize KRAS-GTP primarily by prevent-ing GAP protein-stimulated GTP hydrolysis, thereby stabilizingthe active form of KRAS. We report that the tumor suppressorNF1 is comutated in KRAS G13-mutated cells. NF1 GAP proteinis inactive against KRAS G12 and Q61-mutated KRAS butstimulates GTP hydrolysis when bound to KRAS G13D. KRASG13D mutant cells also respond to EGFR inhibitors in aneurofibromin-dependent manner. Crystallographic analysis ofwild-type and G13D KRAS complexed with neurofibrominprovides the structural basis for neurofibromin-mediated GTPhydrolysis. This suggests that neurofibromin-competent KRASG13-mutated cancers may respond to EGFR therapies.

Author contributions: D.R., T.H.T., and S.D. performed research; D.R., T.H.T., S.D., R.M.S.,F.M., D.K.S., and M.H. analyzed data; F.M., D.K.S., and M.H. supervised the project; andD.R., T.H.T., R.M.S., F.M., D.K.S., and M.H. wrote the paper.

Reviewers: K.C., Brigham andWomen’s Hospital and Harvard Medical School; and K.D.W.,The University of Texas Southwestern Medical Center.

Competing interest statement: F.M. is a consultant for the following companies: AduroBiotech, Amgen, Daiichi Ltd., Ideaya Biosciences, Kura Oncology, Leidos Biomedical Re-search, Inc., PellePharm, Pfizer Inc., PMV Pharma, Portola Pharmaceuticals, and QuantaTherapeutics; has received research grants from Daiichi Ltd. and is a recipient of fundedresearch from Gilead Sciences; is a consultant and cofounder for the following companies(with ownership interest including stock options): BridgeBio Pharma, DNAtrix Inc., OlemaPharmaceuticals, Inc., and Quartz; and is scientific director of the NCI RAS Initiative at theFrederick National Laboratory for Cancer Research/Leidos Biomedical Research, Inc. M.H.is an employee of Quanta Therapeutics.

Published under the PNAS license.

Data deposition: The atomic coordinates and structure factors of the wild-type KRAS–NF1GRD and KRASG13D–NF1GRD complexes reported in this paper have been deposited inthe Protein Data Bank, www.pdb.org (PDB ID codes 6OB2 and 6OB3, respectively).1D.R. and T.H.T. contributed equally to this work.2To whom correspondence may be addressed. Email: frank.mccormick@ucsf.edu,dhirendra.simanshu@nih.gov, or matthew.holderfield@quantatx.com.

3Present address: Quanta Therapeutics, Research Division, San Francisco, CA 94158.

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1908353116/-/DCSupplemental.

First published October 14, 2019.

22122–22131 | PNAS | October 29, 2019 | vol. 116 | no. 44 www.pnas.org/cgi/doi/10.1073/pnas.1908353116

Dow

nloa

ded

by g

uest

on

Apr

il 2,

202

1

http://crossmark.crossref.org/dialog/?doi=10.1073/pnas.1908353116&domain=pdfhttps://www.pnas.org/site/aboutpnas/licenses.xhtmlhttp://www.pdb.orghttp://www.rcsb.org/pdb/explore/explore.do?structureId=6OB2http://www.rcsb.org/pdb/explore/explore.do?structureId=6OB3mailto:frank.mccormick@ucsf.edumailto:dhirendra.simanshu@nih.govmailto:matthew.holderfield@quantatx.comhttps://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1908353116/-/DCSupplementalhttps://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1908353116/-/DCSupplementalhttps://www.pnas.org/cgi/doi/10.1073/pnas.1908353116

pancreas. This may be due in part to allelic differences in signaltransduction (10–12). It has also been proposed that KRASG13D colorectal cancers may respond to inhibition of upstreamsignaling (13, 14). Epidermal growth factor receptor (EGFR)inhibitors are approved for treatment of KRAS wild-type (WT)colorectal cancers, but KRAS mutations are contraindicated. Ret-rospective analyses of the EGFR inhibitor cetuximab showed amodest response in KRAS G13D-mutated colorectal cancers whileKRAS G12-mutated CRCs were resistant. However, these resultswere not observed for panitumumab (15), and a subsequent clinicaltrial showed no response in KRAS G13D-mutated CRC patientstreated with cetuximab as a single agent or in combination withirinotecan (16). Clearly, the presence of the KRAS G13D allelealone is not sufficient to identify KRAS-mutated tumors that mayrespond to anti-EGFR therapies. However, CRC tumor genomesare frequently unstable and diverse (17). The complex underlyinggenetics and biology of KRAS-mutated CRCs raise the possibilitythat a subset of KRAS G13D-mutated colorectal cancers may beEGFR-dependent, and identifying a functional biomarker may helpto further stratify patients.Here we report increased comutation of the RasGAP NF1 with

KRAS G13-mutated cancers and show that KRAS G13D-mutatedcells can respond to EGFR inhibitors in an NF1-dependent man-ner. In addition, we show how NF1 influences KRAS GTP levels in

KRAS G13 mutant cells and that NF1 hydrolyzes GTP bound toKRAS G13D-mutated oncoproteins in biochemical assays. We alsoshow structures of the wild type and G13D mutant of KRAS incomplex with the catalytic region (GRD) of NF1. Since thesestructures were obtained using KRAS bound to GMPPNP (anonhydrolyzable analog of GTP), they represent a ground-stateconformation of the RAS–NF1 complex. Structural comparisonwith the HRAS–RASA1GRD complex obtained in the transi-tion state provides insights into conformational changes thatoccur when RAS–RasGAP complexes move from the ground stateto the transition state. Structural analysis of the KRASG13D–NF1GRD complex provides a rationale for a significant level ofGAP-mediated GTP hydrolysis in this mutant compared with G12and Q61 mutants.

ResultsMutational Landscape of KRAS G13-Mutated Cancers. It has beenreported that gastric tumor types are associated with high overallmutational burdens (17). To determine if KRAS G13D-mutatedcancers appear in a distinct genetic background compared withother KRAS-mutated cancers, a systematic class comparison wasperformed between KRAS G12 mutant (G12x) samples andKRAS G13 mutant (G13x) samples using The Cancer GenomeAtlas (TCGA) data (Fig. 1A). Only sample sets with more than

Fig. 1. KRAS G13D CRC cells are sensitive to NF1-mediated GTP hydrolysis. (A) Total number of mutations per sample in KRAS G12- and G13-mutated COADfrom TCGA cancer samples (P = 0.018). (B) Mutational burden comparing KRAS codon 12 and codon 13 mutations with mismatch repair mutations andmicrosatellite instability (MSI-H, MSI high; MSI-L, MSI low; MSS, MS stable) in TCGA COAD samples. (C) Comparison of NF1 mutations and KRAS mutationstatus in CCLE cell lines. Category sample sizes are indicated above each bar. (D) NF1 overexpression in HCT-116, a KRAS G13D/NF1-mutated cell line, leads toreduced KRAS-GTP levels.

Rabara et al. PNAS | October 29, 2019 | vol. 116 | no. 44 | 22123

APP

LIED

BIOLO

GICAL

SCIENCE

S

Dow

nloa

ded

by g

uest

on

Apr

il 2,

202

1

10 G13-mutated samples were included in the analysis, whichincluded colon adenocarcinoma (COAD), stomach adenocarci-noma (STAD), and rectal adenocarcinoma (READ) samples.We found that G13x-harboring COAD samples had a signifi-cantly (t test, P = 0.018; Fig. 1A) higher number of mutationswhen compared with KRAS G12x samples. Increased mutationalburden for the G13x STAD samples did not reach statisticalsignificance (t test; P = 0.053; SI Appendix, Fig. S1A) and nodifference was observed for READ (SI Appendix, Fig. S1B). Highmutation rate in colorectal tumors is often associated with bothmicrosatellite instability (MSI) and loss of mismatch repair(MMR) gene expression. Therefore, we also compared theseclinical terms between the 2 mutant classes. For COAD, therewas a significant difference (P = 0.014) between the G13x andG12x mutant classes for the clinical term associated withmicrosatellite instability (Fig. 1B). Loss of MMR and increasedMSI was also observed for G13x STAD samples, but both termsfailed to reach statistical significance (MSI: P = 0.095) and nodifference was observed for G13x vs. G12x READ samples (SIAppendix, Fig. S1 C and D).These observations led us to look more closely at the muta-

tional spectra present in the G12x and G13x sample classes. Ascreen to query the top 100 mutated genes (collapsing all per-gene mutations for each sample into a single instance) wasperformed in each sample class and then evaluated for differ-ences in mutational frequencies between the classes. All of theidentified statistically significant differentially mutated genesshowed an enrichment in mutation in the G13-mutated sampleclass relative to the G12-mutated sample class (as this classshows significantly higher mutational frequencies) (SI Appendix,Fig. S1 E–G). This pattern was observed for COAD, STAD, andREAD samples. Interestingly, there were several shared differ-entially mutated genes between the COAD and STAD tumorgroups (PCLO, KMT2D, MBD6, CACNA1E, MDN1, ANK3, andZDBF2). Importantly, among the genes showing mutationalenrichment in the G13x KRAS mutant samples in COAD, therewere several receptor tyrosine kinases—FGFR2, EGFR, INSR,and ALK. Lists of statistically significant differentially mutatedgenes identified for each tumor type are provided in SI Appendix.Cell-line data were also used to perform a mutational en-

richment analysis. Mutational frequencies of genes within theRAS pathway were compared among G12x and G13x KRASmutant cell lines using the COSMIC cell line collection. En-richment analysis revealed that ARHGAP35, CDC25A, CNKSR1,DAB2IP1, EIF4EBP1, IRS2, PAK3, PRKAA1, RASA3, SPRED2,and NF1 were enriched in KRAS G13x-mutated cell lines. NF1was the most common comutation identified, and many of thecell lines with multiple comutations frequently carried NF1mutations as well (Dataset S1).

Comutational Frequencies of NF1 with Different KRAS Mutant Alleles.To confirm that observation, the frequency of NF1 mutationswas determined across several KRAS mutant alleles. Silent NF1mutations were removed for this analysis, although the resultsare still significant if synonymous point mutations are included.Indeed, the G13-mutated samples showed a higher comutationfrequency of NF1 than the G12-mutated cell lines (0.5 vs. 0.06;Fig. 1C). Interestingly, the comutational frequency in the Q61mutant class was zero, while the A146 mutant samples were el-evated similar to the G13-mutated class (Fig. 1C). Together,these observations are consistent with a model whereby both theKRAS G12 and Q61 mutant alleles are able to operate in-dependent of NF1 mutational status while both G13 and A146mutations appear to benefit from an NF1 comutation. Finally,the cBioPortal was used to identify KRAS and NF1 comutationstatus in a larger sample size of patient tumor samples from 155genome-sequencing studies across multiple cancer indications(18, 19). Consistent with the enriched comutations found in the

COSMIC cancer cell lines, ∼12% of KRAS G13x-mutated sam-ples also have NF1 mutations (n = 371), compared with 5.5% ofNF1 comutations in KRAS G12x-mutated samples (n = 2,695).

KRAS G13D Is Sensitive to NF1-Stimulated GTP Hydrolysis. To test ifNF1 expression functionally alters KRAS activity in a KRASG13D CRC cell line, HCT-116 cells were transfected with thecatalytic domain of NF1, which binds to the KRAS WT andcatalyzes GTP hydrolysis. Cellular levels of KRAS GTP werereduced after NF1 expression, indicating that NF1 is functionallyactive and promotes GTP hydrolysis in a KRAS G13D-mutatedCRC cell line, and KRAS G13D protein may be sensitive toNF1-stimulated GTP hydrolysis (Fig. 1D).The ability of NF1 to catalyze GTP hydrolysis directly in

complex with KRAS G13D was then tested in vitro using puri-fied KRAS and NF1 proteins. Consistent with previously pub-lished data (11, 20), RASA1 GAP is active only against wild-typeKRAS, and KRAS proteins with activating mutations at codons12, 13, and 61 are unaffected by RASA1 GAP. RASA1 GAPaccelerated GTP hydrolysis of wild-type KRAS GTP to ∼100-fold above the reported rate of intrinsic GTP hydrolysis for wild-type KRAS, whereas RASA1 GAP did not significantly alter therate of KRAS G12D or G13D (Fig. 2A and Table 1) (11, 21).Similar to RASA1, NF1 stimulated GTP hydrolysis for wild-typeKRAS and did not affect GTP hydrolysis for KRAS G12D (Fig.2B). However, unlike RASA1, NF1 stimulated GTP hydrolysis ofKRAS G13D to a rate much closer to that of NF1 with wild-typeKRAS (Fig. 2B). The ability of NF1 to alter KRAS signal trans-duction was also confirmed in cells. KRAS WT, G13D, or G12Dconstructs were cotransfected with NF1 or GFP control plasmids.MEK phosphorylation levels were measured as an indicator ofMAPK pathway activity because MEK is the first kinase activated inthe MAPK pathway, and may be less prone to attenuation throughDUSP-mediated negative feedback mechanisms that target pERK(22, 23). All KRAS constructs are capable of activating the MAPKpathway leading to elevated pMEK. However, pMEK levels weremarkedly reduced for KRAS wild type and KRAS G13D whencotransfected with NF1 but unaltered in cells cotransfected withKRAS G12D and NF1, suggesting that NF1 inhibits KRAS signaltransduction in both KRAS wild type- and KRAS G13D-expressingcells but NF1 has no effect on cells expressing KRAS G12D (Fig.2C). Together, these results suggest that NF1 directly regulatesKRAS GTP levels in KRAS G13D-mutated cells due to a uniquestructural conformation for KRAS G13D that distinguishes it fromother KRAS oncogenic mutations at codons 12 or 61.

Crystal Structure of Wild-Type KRAS in Complex with NF1GRD. Tobetter understand the underlying mechanism of hydrolysis ob-served for KRAS G13D proteins, we sought to determine thestructures of the wild type and G13D mutant of KRAS incomplex with NF1. NF1 is a large protein comprising 2,818residues and the GAP-related domain is located between resi-dues 1198 and 1530 (Fig. 2D). Although the apo structure ofNF1GRD(GAP333) (residues 1198 to 1530) was solved more than2 decades ago (5), no structural information is available on howNF1 interacts with RAS and carries out NF1-mediated GTPhydrolysis. Despite extensive efforts, we could not obtain crystalsof KRAS in complex with NF1GRD(GAP333). Previous studieshave shown that the central part of NF1GRD is sufficient to carryout GAP-mediated GTPase activity (24). Therefore, a smallerNF1GRD construct was made (GAP255; residues 1209 to 1463)containing the minimum catalytic region by excluding C- and N-terminal regions that have been shown to form the extra domain(Fig. 2D). Using an isothermal titration calorimetry (ITC) ex-periment, NF1-GAP255 showed similar binding affinity toKRAS when compared with NF1-GAP333 (SI Appendix, Fig. S2A and B). Crystallization of NF1GRD(GAP255) in complex withGMPPNP-bound wild-type KRAS (hereafter referred to as the

22124 | www.pnas.org/cgi/doi/10.1073/pnas.1908353116 Rabara et al.

Dow

nloa

ded

by g

uest

on

Apr

il 2,

202

1

https://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1908353116/-/DCSupplementalhttps://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1908353116/-/DCSupplementalhttps://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1908353116/-/DCSupplementalhttps://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1908353116/-/DCSupplementalhttps://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1908353116/-/DCSupplementalhttps://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1908353116/-/DCSupplementalhttps://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1908353116/-/DCSupplementalhttps://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1908353116/-/DCSupplementalhttps://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1908353116/-/DCSupplementalhttps://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1908353116/-/DCSupplementalhttps://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1908353116/-/DCSupplementalhttps://www.pnas.org/cgi/doi/10.1073/pnas.1908353116

KRAS–NF1GRD complex) was successful and we solved thestructure of this protein–protein complex at 2.85 Å (SI Appendix,Table S1). The domain organization showing the location ofGAP255 in the NF1 protein and the overall structure of KRAS–NF1GRD are shown in Fig. 2 D and E, respectively. Despite hav-ing a relatively low amino acid sequence identity of 30% betweenthe GRD of RASA1 and NF1, they form a similar tertiary structure(with an rmsd of 1.8 Å for Cα atoms) and interact with KRASin a similar way (Fig. 2F). In the KRAS–NF1GRD structure,NF1GRD forms a crescent-shaped structure and KRAS binds inthe central groove via residues of the switch regions.Measurement of binding affinity using ITC agrees with previous

observations that NF1GRD binds to RAS with high affinity (SI Ap-pendix, Fig. S2A) (20). Previous studies have shown that RASA1binds to RAS with a 50-fold lower affinity than NF1 (20, 25, 26).Structural analysis of the protein–protein interface in previouslysolved HRAS–RASA1GRD and KRAS–NF1GRD complexes de-scribed here shows similar buried surface areas and hydrogen-bonding patterns between these 2 complexes (SI Appendix, Fig.S3). At the RAS–RasGAP interface, nearly half the residues that

form the RasGAP interface are different between NF1 and RASA1(SI Appendix, Fig. S4). However, the structural comparison showedthe presence of 4 strong salt-bridge interactions in the KRAS–NF1GRD complex spanning the protein–protein interface whereasthe HRAS–RASA1GRD complex has only 2 such interactions at itsinterface (SI Appendix, Fig. S3 C and D). It is likely that the addi-tional salt bridges present at the KRAS–NF1GRD interface providea higher-affinity interaction of KRAS with NF1 compared withRASA1 GAP.

NF1-Mediated GTP Hydrolysis in KRAS and Conformational Changes inRAS–RasGAP Complexes. The structure of GMPPNP-bound wild-type KRAS complexed with NF1GRD shows the ground-stateconformation with the switch I region in the active conforma-tion. Despite our extensive efforts, we could not obtain well-diffracting crystals of KRAS–NF1GRD in the transition state(using GDP and AlF3). Thus, we compared our ground-statestructure with the previously reported transition-state structure ofthe HRAS–RASA1GRD complex to gain insights into NF1-mediatedGTP hydrolysis and conformational changes that occur when the

Fig. 2. RASA1 and NF1 GAP-mediated GTP hydrolysis with multiple KRAS oncoproteins. Intrinsic and GAP-stimulated KRAS GTPase activity was measuredusing EnzChek. (A) RASA1-stimulated hydrolysis of KRAS WT/G12D/G13D (represents 1 EnzChek assay with 3 replicates of each KRAS mutant all on the sameplate, error bars represent standard deviation). (B) NF1-stimulated hydrolysis of KRAS WT/G12D/G13D (represents 1 EnzChek assay with 6 replicates of eachKRAS mutant all on the same plate, error bars represent standard deviation). (C) Decreased pMEK levels in WT and G13D when cotransfected with FLAG-KRASWT or G13D and NF1 in NCI-H1355 cells. (D) Domain organization of human NF1. (E) The overall structure of GMPPNP-bound wild-type KRAS in complex withNF1GRD. KRAS and NF1GRD are shown in cartoon representation and colored pink and green, respectively. GMPPNP and the arginine finger (NF11276) areshown in stick representation where oxygen, nitrogen, and phosphorus atoms are colored red, blue, and orange, respectively. (F) Structural comparison of theground-state KRAS–NF1GRD complex with the transition-state HRAS–RASA1GRD complex. Here KRAS and NF1GRD are color-coded as in E whereas HRAS andRASA1GRD shown in cartoon representation are colored orange. Secondary structural elements around the arginine finger show rotational displacement ofGAP-related domains during the conversion from the ground state to the transition state.

Table 1. Biochemical comparison of KRAS WT and oncogenic mutants

KRASRASA1-stimulated rate

of hydrolysis kobs 10−5, s−1

NF1-stimulated rateof hydrolysis kobs 10

−5, s−1NF1 binding

affinity KD, μM

WT 2599.00 ± 398.46 5027.80 ± 2241.63 1.2 ± 0.2G12D ND 194.70 ± 85.42 8.1 ± 0.4G13D ND 3457.00 ± 1789.87 6.4 ± 0.7G13C ND 3098.60 ± 1024.47 11.0 ± 0.0Q61R ND ND 18.4 ± 4.4

The hydrolysis rates listed are derived from averaging the hydrolysis rates of each KRAS mutant from multipleEnzChek hydrolysis runs. ND, not determined (hydrolysis activity too low to measure).

Rabara et al. PNAS | October 29, 2019 | vol. 116 | no. 44 | 22125

APP

LIED

BIOLO

GICAL

SCIENCE

S

Dow

nloa

ded

by g

uest

on

Apr

il 2,

202

1

https://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1908353116/-/DCSupplementalhttps://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1908353116/-/DCSupplementalhttps://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1908353116/-/DCSupplementalhttps://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1908353116/-/DCSupplementalhttps://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1908353116/-/DCSupplementalhttps://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1908353116/-/DCSupplementalhttps://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1908353116/-/DCSupplementalhttps://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1908353116/-/DCSupplemental

RAS–RasGAP complex shifts from the ground state to the transitionstate (4). The structural superposition of HRAS–RASA1GRD andKRAS–NF1GRD using RAS proteins shows an ∼8° rotation and1.8-Å translation of the GAP domain toward RAS when it movesfrom the ground state to the transition state (Fig. 3A). This rotationbrings the GAP protein close to RAS and places the arginine fingerin the active site, where it stabilizes and orients KRASQ61 for aninline nucleophilic attack on the gamma-phosphate of GTP. Calcu-lation of the interaction interface shows an increase of 500 Å2 forboth KRAS and NF1GRD as they shift from the ground state to thetransition state. Similar rotation, although to a large extent (∼20°),has been seen in the case of RHO proteins, where the structures ofRHO–RhoGAP complexes are available in both ground and transi-tion states (27).In the ground-state structure of KRAS–NF1GRD, the side-chain

atoms of KRASY32 occupy the position of the arginine finger(NF1R1276) as seen in the transition-state structure of the HRAS–RASA1GRD complex (Fig. 3 B–D). Side-chain atoms of KRASY32

and KRASQ61 interact with each other via a bridging water mol-ecule. Unlike the transition-state structure, where HRASQ61 formsa hydrogen bond with the main-chain carbonyl oxygen of the ar-ginine finger, in the ground-state structure KRASQ61 forms ahydrogen bond with the main-chain amide nitrogen of NF1G1277, aresidue following the arginine finger (Fig. 3D). Structural super-position of ground- and transition-state structures suggests that tofacilitate NF1-mediated GTPase activity in wild-type KRAS, a ro-tation of NF1 toward KRAS would allow placement of the argininefinger in the nucleotide-binding pocket by swapping its position withKRASY32. This rotation would now bring NF1R1276 (instead ofNF1G1277) present in the finger loop close to KRASQ61, allowing anew hydrogen bond to form between the amide nitrogen of theKRASQ61 side chain and main-chain carbonyl oxygen of the arginine

finger. These conformational changes in the active site wouldstabilize and orient KRASQ61 and NF1R1276 in the active site forthe GTP hydrolysis reaction via formation of the transition state,as observed in the structure of the HRAS–RASA1GRD complex.

Structure of KRASG13D Complexed with NF1GRD Provides a Rationalefor NF1-Mediated GTPase Activity.As predicted earlier, unlike wild-type KRAS, the oncogenic mutants G12D, G13D, and Q61R ofKRAS show relatively weak affinity for NF1GRD. Among these 3mutants, G13D and G12D mutants of KRAS bind toNF1GRD with a relatively higher affinity than Q61R mutants(Table 1 and SI Appendix, Fig. S2 C–F). As mentioned above,the rate of NF1-mediated GTP hydrolysis for KRASG13D issurprisingly similar to wild-type KRAS (Table 1). To understand thestructural basis of NF1-mediated GTPase activity in the G13Dmutant of KRAS, we solved the structure of GMPPNP-boundKRASG13D complexed with NF1GRD(GAP255) at 2.1 Å (SI Ap-pendix, Table S1). The overall structure and the interaction in-terface of the KRASG13D–NF1GRD complex are similar tothose of the wild-type KRAS–NF1GRD structure (Fig. 4A). Inthe KRASG13D–NF1GRD structure, the presence of side-chainatoms of KRASD13 displaces KRASY32 away from the nucleotide-binding pocket (SI Appendix, Fig. S5) and this results in the dis-placement of the finger loop containing the arginine finger andsurrounding residues in NF1GRD at the KRAS–NF1 interface.This vertical displacement, which is close to 4 Å for the Cα atom ofNF11276, allows partial entry of the arginine finger into thenucleotide-binding pocket, where it interacts with KRASD13

and occupies the position of KRASY32 observed in the wild-typeKRAS–NF1GRD structure (Fig. 4B). Unlike the KRAS–NF1GRD

structure, one of the side-chain carbonyl oxygen atoms ofKRASD13 of the KRASG13D–NF1GRD structure interacts with

Fig. 3. Structural analysis of the active-site pocket and protein–protein interaction interface in the KRAS–NF1GRD and HRAS–RASA1GRD complexes. (A)Conformational changes in NF1GRD during the conversion from the ground state to the transition state. (B) Schematic of the GAP-mediated GTP hydrolysisreaction in RAS. (C) Active-site pocket in the KRAS–NF1GRD complex. (D) Structural superposition of the active-site pocket in the KRAS–NF1GRD and HRAS–RASA1GRD complexes.

22126 | www.pnas.org/cgi/doi/10.1073/pnas.1908353116 Rabara et al.

Dow

nloa

ded

by g

uest

on

Apr

il 2,

202

1

https://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1908353116/-/DCSupplementalhttps://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1908353116/-/DCSupplementalhttps://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1908353116/-/DCSupplementalhttps://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1908353116/-/DCSupplementalhttps://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1908353116/-/DCSupplementalhttps://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1908353116/-/DCSupplementalhttps://www.pnas.org/cgi/doi/10.1073/pnas.1908353116

the main-chain amide hydrogen of NF1G1277, and the side chain ofKRASQ61 forms a hydrogen bond with the side chain of NF1R1391(Fig. 4C). The local conformational changes observed in the fingerloop of the KRASG13D–NF1GRD structure suggest that it has in-herent flexibility that could allow it to place the arginine finger inthe active-site pocket, even in the case of minor steric hindrancedue to the G13 mutation.Structural comparison of the KRASG13D–NF1GRD complex

with the KRAS–NF1GRD and HRAS–RASA1GRD complexessuggests that to enable NF1-mediated GTPase activity inKRASG13D, besides the above-mentioned rotation and translationof NF1GRD toward KRASG13D, it would require that KRASD13

adopts another rotamer conformation which points away from thegamma-phosphate so that the arginine finger can occupy its placeand interact with phosphate moieties of GTP. Examination ofother possible side-chain rotamer conformations of KRASD13 inthe KRASG13D–NF1GRD complex structure reveals a plausiblerotamer conformation of the KRASD13 side chain pointing towardKRASK117 (away from the gamma-phosphate) and forming a salt-bridge interaction with it. Thus, it is likely that the conversion ofthe ground state to the transition state would be accompanied by achange to this rotamer conformation of the KRASD13 side chainin the KRASG13D–NF1GRD complex. This would allow the argi-nine finger to enter the active-site pocket and orient KRASQ61 asin the wild-type KRAS complex for inline nucleophilic attack onthe gamma-phosphate of GTP. These conformational changesincorporated in the modeled structure of the KRASG13D–NF1GRD complex in the transition state provide a structural basisfor the observed NF1-mediated GTP hydrolysis in this mutant (Fig.4D). This model also predicts that other oncogenic KRAS muta-tions at codon 13 may also be sensitive to NF1-stimulated GTPhydrolysis.

KRAS G13C Is Also Sensitive to NF1-Stimulated GTP Hydrolysis. Todetermine if NF1-mediated GTPase activity is specific for KRASG13D or if other codon 13 mutations are also NF1-sensitive,NF1 biochemical activity was tested with purified KRAS G13Cprotein. Although it only accounts for a fraction of a percent ofhuman cancers, KRAS G13C is the second most common KRAScodon 13 mutation observed in human tumors. As predicted andconsistent with our model, the NF1 GRD stimulated GTP hydrolysisin KRAS G13C with a kobs rate similar to KRAS G13D hydrolysisstimulation by NF1 (Fig. 4E and Table 1). To determine if NF1 canaffect KRAS G13C activity in cells, NF1 was transfected into NCI-H1355 cells, which carry an endogenous KRASG13C mutation. Con-sistent with the in vitro biochemical data, KRAS GTP levels werereduced in the presence of NF1 (Fig. 4F).

NF1 Alters Response to EGFR Inhibitors in KRAS WT and G13D-MutatedCells. It has been suggested that a subset of KRASG13D-mutatedcolorectal cancers may respond to EGFR-inhibitor therapies (13,14). However, clinical responses to cetuximab have not been sta-tistically significant, and no underlying mechanism has been pro-posed. We postulated that NF1 activity may alter drug sensitivityin KRAS G13x-mutated colorectal cancer cells. Because manyKRASG13D-mutated CRC cell lines also harbor an NF1 mutation,we chose to test this in SW48 cells, a CRC cell line with a KRASand NF1 wild-type genetic background. SW48 cells engineered tocarry either a KRASG12D or KRASG13D mutation in the endoge-nous locus were treated with the EGFR inhibitor gefitinib. Theparental KRAS wild-type cells had an EC50 of 7.26 nM, whilethe KRASG12D SW48 cells were resistant up to 10.49 μM. TheKRASG13D SW48 cells had a moderate sensitivity to gefitinib with anEC50 of 35.54 nM (Fig. 5A). Similar results were obtained witherlotinib, another small-molecule EGFR inhibitor (SI Appendix,

Fig. 4. RASA1 and NF1 GAP-mediated GTP hydrolysis with multiple KRAS oncoproteins. (A) Structural superposition of NF1(GAP255) complexed with WTKRAS and KRASG13D. (B) Enlarged view of the active-site pocket in the superposed structures of NF1(GAP255) complexed with WT KRAS and KRASG13D (groundstate). (C) The active-site pocket in the crystal structure of the KRASG13D–NF1GRD complex in the ground state. (D) The active-site pocket in the modeledstructure of the KRASG13D–NF1GRD complex in the transition state. (E) GTP hydrolysis of KRAS G13C mediated by RasGAP proteins (error bars representstandard deviation of 3 replicates). (F) KRAS-GTP levels after NF1 expression in NCI-H1355, a KRAS G13C-mutated CRC cell line.

Rabara et al. PNAS | October 29, 2019 | vol. 116 | no. 44 | 22127

APP

LIED

BIOLO

GICAL

SCIENCE

S

Dow

nloa

ded

by g

uest

on

Apr

il 2,

202

1

https://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1908353116/-/DCSupplemental

Fig. S6). In addition, the SW48 isogenic cell lines had comparabledoubling times while maintaining the cells in culture, and thus re-sponse to EGFR inhibitors is unlikely to be caused by differences inrates of proliferation. To determine if EGFR-inhibitor sensitivity inthe KRASG13D SW48 cells is NF1-dependent, NF1 expression wasdepleted using small interfering RNA (siRNA) transfection. mRNAlevels were depleted by 70 to 80% and the NF1 protein was notdetectable by Western blot (Fig. 5D and SI Appendix, Fig. S7). AfterNF1 depletion, KRASG13D SW48 cells became resistant to gefitinib,whereas KRASG12D SW48 cells remained resistant to gefitinibwithout NF1 (Fig. 5 B and C). Consistent with these data, Westernblots showed reduced pMEK levels after gefitinib treatment in theparental KRAS wild-type SW48 cells, while pMEK levels remainedunchanged in the KRASG12D SW48 cells even at the highest con-centrations tested. NF1 knockdown partially restored pMEK levelsin the KRAS wild-type and G13D SW48 cells but remained un-changed in the KRASG12D SW48 cells (Fig. 5D). Together, thesedata suggest that, in an isogenic system, the presence of aKRASG13D mutation may confer sensitivity to EGFR inhibitors,but only in the context of functional NF1 protein.

DiscussionKRAS residues G12, G13, and Q61 are all important for co-ordinating GTP binding, and mutation at these residues has beensuggested to disrupt GAP-stimulated GTP hydrolysis and con-stitutively activate the RAF/MAPK pathway (28). Mutations atthese sites occur at different frequencies in different cancer in-dications, and KRASG13D mutations appear almost exclusively in

gastrointestinal cancers (16, 28–30). This observation raises theintriguing possibility that although all KRAS mutations stabilizeKRAS GTP levels, the different KRAS oncogenic alleles mayhave distinct biological functions. Indeed, the biochemistry ofKRAS proteins harboring G13 mutations appears distinct fromcodon 12 and 61 variants. KRASG13-mutated proteins have beendescribed with reduced but measurable GAP-stimulated GTPaseactivity, suggesting that KRASG13x may be less oncogenic thanKRASG12x or KRASQ61x cells (11). Here we report that withingastric cancers, tumor genomes with a KRASG13x mutation areassociated with higher genetic instability, and frequently exhibitcomutations within the RAS pathway. Furthermore, ∼12% ofKRASG13x-mutated patient samples and 50% of KRASG13x-mutated cell lines also carry an NF1 comutation (Fig. 1C) whileNF1 mutations appear mutually exclusive with KRASG12x- andKRASQ61x-mutated cells, suggesting that KRASG13x cells maybenefit from an additional mutation to fully activate the RASpathway. Interestingly, NF1 was also comutated in KRASA146T-mutated tumor samples and cell lines (Fig. 1C), again suggestingthat low-frequency, weakly oncogenic KRAS-mutated cells mayrequire an additional comutation in the RAS signal-transductionpathway.Although all oncogenic KRAS alleles appear resistant to

GTP hydrolysis by RASA1 GAP (11), the 2 most commoncodon 13 variants, KRASG13D and KRASG13C, are sensitive toNF1-stimulated GTPase activity in vitro and in cells. Restorationof NF1 activity in NF1-mutated cells stabilizes KRASG13D GTPlevels and activates downstream signal transduction. Crystal

Fig. 5. KRAS G13D is sensitive to EGFR inhibition in an NF1-dependent manner. SW48 isogenic cell line dose–response to gefitinib. (A) Dose–response of WT,G12D, and G13D isogenic cell lines without siRNA treatment (error bars represent standard deviation). (B) G12D siRNA treated shows no change in response togefitinib (error bars represent standard deviation). (C) G13D shows resistance to drug when NF1 is knocked down (error bars represent standard deviation).(D) G13D shows pMEK resistance when NF1 is knocked down, whereas the WT shows sensitivity with or without NF1 and G12D remains resistant in bothsituations. Drug treatment was for 48 h without EGF stimulation.

22128 | www.pnas.org/cgi/doi/10.1073/pnas.1908353116 Rabara et al.

Dow

nloa

ded

by g

uest

on

Apr

il 2,

202

1

https://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1908353116/-/DCSupplementalhttps://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1908353116/-/DCSupplementalhttps://www.pnas.org/cgi/doi/10.1073/pnas.1908353116

structures of the wild type and G13D mutant of GMPPNP-boundKRAS in complex with the GAP-related domain of NF1 describedhere provide the structural basis of NF1-mediated GTP hydrolysisin the wild type as well as G13 mutants of KRAS. Structuralcomparison between the KRAS–NF1GRD complex and HRAS–RASA1GRD complex suggests that, like the Rho–RhoGAP com-plex, a rotation of 8° of NF1GRD would enable conversion from theground state to the transition state, resulting in placement of thecatalytic arginine finger inside the KRAS active site. Structural re-sults described here validate that residue R1276 acts as an argininefinger in the NF1-mediated GTP hydrolysis reaction.Comparison of the wild type and G13D mutant of KRAS

(GMPPNP-bound) in complex with NF1GRD suggests that G13D/Cmutants can allow placement of the arginine finger (present in theflexible finger loop) in the active-site pocket by adopting a side-chain rotamer conformation, which points away from the gamma-phosphate. In the case of KRASG13D, interaction of D13 with K117is likely to stabilize its alternate rotamer conformation and preventany steric clash between D13 (KRAS) and the arginine finger(NF1). Previously, it has been suggested that unlike G12, oncogenicmutation at G13 in KRAS is less stringent in the steric requirementsand is able to tolerate small perturbations, and thus can allowplacement of the arginine finger in the catalytic pocket for NF1-mediated GTPase activity (31). As suggested, loss of NF1-mediatedGTP hydrolysis in G12 mutants of KRAS is likely due to sterichindrance, which does not allow placement of the arginine finger inthe nucleotide-binding pocket for any possible rotamer conforma-tions the mutated residue at the G12 position adopts. Insensitivityof KRAS Q61 mutants to NF1 and RASA1GAPs is likely due to itsinability to stabilize the arginine finger and orient the catalytic waterfor nucleophilic attack.As suggested previously, formation of RAS–RasGAP complexes

in the presence of a nonhydrolyzable GTP analog represents aground-state conformation, whereas formation of RAS–RasGAPcomplexes in the presence of GDP and aluminum fluoride mimicsthe transition state of GAP-mediated GTPase activity (31). Previousbiochemical studies have shown that G13A and G13S mutants ofHRAS bind to NF1 with close to wild-type HRAS affinity and forman HRAS–NF1 complex in the presence of aluminum fluoride,suggesting NF1-mediated GTPase activity in these mutants (31).Unlike G13A/S, G13R showed weaker affinity and a reduced levelof complex formation in the presence of aluminum fluoride. Sur-prisingly, in that study, no interaction was observed between theG13D mutant of HRAS and NF1GRD (31). We measured theaffinity of KRAS mutants and NF1GRD using isothermal titrationcalorimetry, and the dissociation constants (Kds) of G13D/Cmutants were close to that of wild-type KRAS. Biochemical andstructural data presented here suggest that the mutation of G13in RAS to other amino acids with a relatively small side chain(Ala, Ser, Cys, or Asp) is likely to be sensitive to NF1, and thissensitivity would decrease when G13 is mutated to a larger side-chain residue such as arginine.Interestingly, unlike NF1, G13D/C mutations in KRAS are

insensitive to a RASA1-mediated GTP hydrolysis reaction (Figs.2A and 4E). Analysis of biochemical and structural differencesbetween NF1 and RASA1 interactions with RAS provides apossible rationale for this observation. Previous studies haveshown that the binding affinity of RAS for NF1GRD is 50-foldhigher than RASA1GRD (20, 25, 26). The association and dis-sociation rate constants for RASA1 have been shown to be atleast 2- and 100-fold faster than NF1 (20) and, because of this, itis difficult to measure binding affinity between KRAS and RASA1using ITC. In NF1, the arginine finger is followed by Gly instead ofAla as in the case of RASA1, and it has been suggested that thisresults in higher flexibility in the finger loop of NF1 (5). It ispossible that relatively weaker affinity and faster kinetics between

KRAS and RASA1 and the relatively rigid finger loop is responsiblefor G13 mutants’ insensitivity toward RASA1.NF1 biochemical activity with KRAS G13-mutated proteins

may also partially explain the prevalence of these mutations incolorectal cancers. The high mutation rate found in gastriccancer patient samples includes functionally redundant mecha-nisms that may allow for less oncogenic KRAS variants toemerge. Comutation of tumor suppressors and signaling modi-fiers within the RAS pathway may augment activity of KRASG13-mutated oncoproteins. It is also important to note thatcolorectal cancers frequently harbor underlying alterations in theWNT and TGF-β pathways, mutations that are rare in otherKRAS-mutated cancer indications but often cross-talk withMAPK signaling. WNT signaling can stabilize oncogenic KRASprotein (32) and TGF-β signaling augments KRAS-driven on-cogenesis in colorectal cancers (33). A better understanding ofthese mechanisms may help explain the high prevalence of KRASG13 mutations in colon cancers specifically and how WNT andTGF-β signaling affect NF1 in KRAS G13-mutated coloncancer cells.Finally, the structural and biochemical properties unique to G13-

mutated KRAS proteins may have significant implications forKRAS-mutated colorectal cancer patients. The clinical observationthat KRAS G13D CRC may respond to EGFR inhibition (13, 14)may be explained by the increased sensitivity to NF1-stimulatedGTP hydrolysis. In the presence of active, functional NF1,KRASG13D cancer cells may have reduced KRAS GTP levels, andbe partially dependent on upstream signaling through EGFR orother receptor tyrosine kinases. Although not tested in this study,G13x-mutated cell lines that also harbor FGFR2-, EGFR-, INSR-,or ALK-activating point mutations may similarly respond to directinhibition of the mutated receptor tyrosine kinase. Conversely,cancers with KRASG12x or KRAS61x mutations are resistant to NF1activity. Codon 12 or 61 mutations are fully active and independentof NF1 mutation status or NF1 activity, and therefore resistant toinhibition of upstream stimuli. Therefore, identification of the ap-propriate biomarker for active NF1 could expand the patient pop-ulation for receptor tyrosine kinase inhibitors to include a subset ofKRASG13x-mutated cancers.

Materials and MethodsBioinformatics. TCGA mutation and clinical data were downloaded from theBroad Institute GDAC firehose website (https://gdac.broadinstitute.org/).Tumor types with more than 10 G13x and G12x mutant samples includedCOAD, READ, and STAD. Per-class nonsynonymous mutation counts werecompared for each tumor type and evaluated for statistical significance us-ing the t test. Clinical term data were evaluated similarly except that theFisher exact test was applied. For cell-line analysis, mutation data weredownloaded from the COSMIC website’s cell-line project area (https://cancer.sanger.ac.uk/cell_lines). From these data, the top 100 mutated genesshowing nonsynonymous comutation with any nonsynonymous KRAS mu-tation were identified and subsequently evaluated for statistical significancecontrasting the G12 and G13 mutant sample classes. Complete lists of thestatistically significant differentially mutated genes identified are providedin SI Appendix. For the NF1 mutational analysis, samples were subdividedinto classes according to their mutational status and their NF1 comutationalfrequency status was determined.

Cell Culture and Transfection. HCT-116 and SW48 isogenic cell lines were li-censed from Horizon Discovery. Both sets of cell lines were cultured in RPMImedium (ATCC; 30-2001) supplemented with 10% fetal bovine serum (FBS)(GE Life Sciences; SH30070.03). NCI-H1355 cells were purchased fromthe ATCC. The cells were cultured in ACL4 media made from DMEM/F12(ATCC; 30-2006), and FBS was used in place of BSA (SI Appendix, Materialsand Methods). SW48 isogenic cell lines were transfected with siRNA usingLipofectamine RNAiMAX (Thermo Fisher Scientific) as previously de-scribed by Vartanian (34). The following sequences were used as theNF1 siRNA: hNF1_558 5′-CUCACUACUAUUUUAAAGAAUA-3′, hNF1_558_as5′-UUCUUUAAAAUAGUAGUGAGGC-3′. Scramble siRNA (Thermo FisherScientific; Ambion Silencer SelectNegative Control 2 siRNA) was used as a

Rabara et al. PNAS | October 29, 2019 | vol. 116 | no. 44 | 22129

APP

LIED

BIOLO

GICAL

SCIENCE

S

Dow

nloa

ded

by g

uest

on

Apr

il 2,

202

1

https://gdac.broadinstitute.org/https://cancer.sanger.ac.uk/cell_lineshttps://cancer.sanger.ac.uk/cell_lineshttps://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1908353116/-/DCSupplementalhttps://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1908353116/-/DCSupplementalhttps://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1908353116/-/DCSupplemental

control in siRNA transfections. HCT-116 cells and NCI-H1355 cells weretransfected with cDNAs using FuGENE HD (Promega) per the manufac-turer’s protocol. The following cDNA constructs were used: HA-NF1-CaaxSTD [CMV51P>HA-TEV-NF1(1198–1509)-CAAX STD] (R711-M06-370), GFP(CMV51p>GFP) (R931-M01-366), KRAS WT (CMV51p>3XFLAG-Hs.KRASopt)(R750-M67-304), KRAS G12D (CMV51p>3xFLAG-HsKRAS4b G12D) (R750-M03-304), and KRAS G13D (CMV51p>3xFLAG-KRAS4b G13D) (R750-M05-304). See SI Appendix, Materials and Methods for sequences and con-struction. HCT-116 and NCI-H1355 cell lines were stimulated by the addition of100 ng/mL EGF for 5 min prior to harvesting for Western blot and active RASpull-down. KRAS GTP levels were measured using the Active Ras Pull-Down andDetection Kit (Thermo Fisher Scientific; 16117) according to the manufacturer’sprotocol. CellTiter-Glo (Promega; G7573) was used for proliferation assays perthe manufacturer’s protocol. Cells were treated with inhibitors (gefitinib; TocrisBioscience; 3000) for 72 h unless otherwise indicated.

Biochemical GTP Hydrolysis Assays. The EnzChek Phosphate Assay Kit (ThermoFisher Scientific; E6646) was used tomeasure GTP hydrolysis in vitro. Reactionswere performed in 384-well microplates with 30 mM Tris (pH 7.5) with 1 mMDTT. Final reaction concentrations in the wells were 200 μM MESG, 5 U/mLPNP, 200 μM GTP, and 10 mM MgCl2. Final concentrations of KRAS rangedfrom 80 to 2.5 μM and final GAP concentrations ranged from 80 to 40 μM.Microplates were read every 12 to 18 s at 360 nm. Data were analyzed bysubtracting the background (no-substrate control) at each measurement.The amount of GTP hydrolyzed (μM) was determined by converting absor-bance to micromolar concentration of GTP hydrolyzed using the linear fit ofthe KH2PO4 standard curve. Data were fit to a 1-phase association curve.Initial enzyme velocity measurements were used to calculate kobs values.

Structure Determination and Analysis. The structure of the wild-type KRAS–NF1GRD complex was solved at 2.85 Å by molecular replacement using theprogram Phaser as implemented in the Phenix/CCP4 suite of programs with aprotein-only version of Protein Data Bank ID codes 3GFT (Q61H-KRAS boundto GMPPNP) and 1NF1 (apo structure of NF1GRD) as search models (35, 36).The initial solution was refined using Phenix.refine and the resulting 2Fo − Fcmap showed clear electron density for the 2 proteins. The model was furtherimproved using iterative cycles of the manual model building in Coot andrefinement using Phenix.refine (35, 37). The structure of GMPPNP-boundKRASG13D complexed with NF1GRD(GAP255) was solved at 2.1 Å using the

structure of the wild-type KRAS–NF1GRD complex. Refinement statistics forthese 2 crystal structures are summarized in SI Appendix, Table S1. Secondarystructural elements were assigned using DSSP (https://swift.cmbi.ru.nl/gv/dssp/).Figures were generated with PyMOL (Schrödinger). Crystallographic andstructural analysis software support is provided by the SBGrid Consortium (38).

Isothermal Titration Calorimetry Measurements. Binding affinities of theGMPPNP-bound wild type and oncogenic mutants of KRAS (1 to 169) withNF1GRD (GAP333 and GAP255) were measured using isothermal titrationcalorimetry. Protein samples were prepared by extensive dialysis in a buffer(filtered and degassed) containing 20 mM Hepes (pH 7.3), 150 mM NaCl,5 mM MgCl2, and 1 mM TCEP. For the ITC experiment, 60 μM KRAS and600 μM NF1-GRD were placed in the cell and syringe, respectively. ITC ex-periments were carried out in a MicroCal PEAQ-ITC instrument (Malvern) at25 °C using an initial 0.4-μL injection and 18 subsequent injections of 2.2 μLeach at 150-s intervals. Data analysis was performed based on a bindingmodel containing “one set of sites” using a nonlinear least-squares algo-rithm incorporated in the MicroCal PEAQ-ITC analysis software (Malvern).

ACKNOWLEDGMENTS. We thank the following members of the ProteinExpression Laboratory (Frederick National Laboratory for Cancer Research)for their help in cloning, protein expression, and purification: AllisonCoward, John-Paul Denson, Dominic Esposito, Randy Gapud, Bill Gillette,Jennifer Mehalko, Simon Messing, Shelley Perkins, Lauren Procter, NityaRamakrishnan, Jose Sanchez Hernandez, Mukul Sherekar, Troy Taylor, andVanessa Wall. We would also like to thank Ming Yi for assistance withproliferation data analysis. We are thankful to the staff at the 24-ID-C/Ebeamlines of the Advanced Photon Source, Argonne National Laboratory,for their help with data collection. This project was funded in whole or inpart with federal funds from National Cancer Institute, NIH ContractHHSN261200800001E. Part of this work is based on research conducted atthe Northeastern Collaborative Access Team beamlines, which are funded bythe National Institute of General Medical Sciences from National Institutes ofHealth Grant P41 GM103403. This research used resources of the AdvancedPhoton Source, a US Department of Energy (DOE) Office of Science UserFacility operated for the DOE Office of Science by Argonne NationalLaboratory under Contract DE-AC02-06CH11357. The content of this publi-cation does not necessarily reflect the views or policies of the Department ofHealth and Human Services, and the mention of trade names, commercialproducts, or organizations does not imply endorsement by the US Government.

1. D. K. Simanshu, D. V. Nissley, F. McCormick, RAS proteins and their regulators in

human disease. Cell 170, 17–33 (2017).2. P. Gideon et al., Mutational and kinetic analyses of the GTPase-activating protein

(GAP)-p21 interaction: The C-terminal domain of GAP is not sufficient for full activity.

Mol. Cell. Biol. 12, 2050–2056 (1992).3. D. Vigil, J. Cherfils, K. L. Rossman, C. J. Der, Ras superfamily GEFs and GAPs: Validated and

tractable targets for cancer therapy? Nat. Rev. Cancer 10, 842–857 (2010).4. K. Scheffzek et al., The Ras-RasGAP complex: Structural basis for GTPase activation

and its loss in oncogenic Ras mutants. Science 277, 333–338 (1997).5. K. Scheffzek et al., Structural analysis of the GAP-related domain from neurofibromin

and its implications. EMBO J. 17, 4313–4327 (1998).6. K. Scheffzek, M. R. Ahmadian, A. Wittinghofer, GTPase-activating proteins: Helping

hands to complement an active site. Trends Biochem. Sci. 23, 257–262 (1998).7. C. Kötting, A. Kallenbach, Y. Suveyzdis, A. Wittinghofer, K. Gerwert, The GAP argi-

nine finger movement into the catalytic site of Ras increases the activation entropy.

Proc. Natl. Acad. Sci. U.S.A. 105, 6260–6265 (2008).8. S. Gysin, M. Salt, A. Young, F. McCormick, Therapeutic strategies for targeting Ras

proteins. Genes Cancer 2, 359–372 (2011).9. B. R. Korf, Malignancy in neurofibromatosis type 1. Oncologist 5, 477–485 (2000).10. G. A. Hobbs, C. J. Der, K. L. Rossman, RAS isoforms and mutations in cancer at a

glance. J. Cell Sci. 129, 1287–1292 (2016).11. J. C. Hunter et al., Biochemical and structural analysis of common cancer-associated

KRAS mutations. Mol. Cancer Res. 13, 1325–1335 (2015).12. N. T. Ihle et al., Effect of KRAS oncogene substitutions on protein behavior: Implications

for signaling and clinical outcome. J. Natl. Cancer Inst. 104, 228–239 (2012).13. W. De Roock et al., Association of KRAS p.G13D mutation with outcome in patients

with chemotherapy-refractory metastatic colorectal cancer treated with cetuximab.

JAMA 304, 1812–1820 (2010).14. S. Tejpar et al., Association of KRAS G13D tumor mutations with outcome in patients

with metastatic colorectal cancer treated with first-line chemotherapy with or with-

out cetuximab. J. Clin. Oncol. 30, 3570–3577 (2012).15. M. Peeters et al., Mutant KRAS codon 12 and 13 alleles in patients with metastatic

colorectal cancer: Assessment as prognostic and predictive biomarkers of response to

panitumumab. J. Clin. Oncol. 31, 759–765 (2013).16. E. Segelov et al., Response to cetuximab with or without irinotecan in patients with

refractory metastatic colorectal cancer harboring the KRAS G13D mutation: Australasian

Gastro-Intestinal Trials Group ICECREAM Study. J. Clin. Oncol. 34, 2258–2264 (2016).

17. M. S. Lawrence et al., Mutational heterogeneity in cancer and the search for newcancer-associated genes. Nature 499, 214–218 (2013).

18. E. Cerami et al., The cBio Cancer Genomics Portal: An open platform for exploringmultidimensional cancer genomics data. Cancer Discov. 2, 401–404 (2012).

19. J. Gao et al., Integrative analysis of complex cancer genomics and clinical profilesusing the cBioPortal. Sci. Signal. 6, pl1 (2013).

20. M. R. Ahmadian, U. Hoffmann, R. S. Goody, A. Wittinghofer, Individual rate constantsfor the interaction of Ras proteins with GTPase-activating proteins determined byfluorescence spectroscopy. Biochemistry 36, 4535–4541 (1997).

21. G. Bollag et al., Biochemical characterization of a novel KRAS insertion mutation froma human leukemia. J. Biol. Chem. 271, 32491–32494 (1996).

22. A. Kucharska, L. K. Rushworth, C. Staples, N. A. Morrice, S. M. Keyse, Regulation of theinducible nuclear dual-specificity phosphatase DUSP5 by ERK MAPK. Cell. Signal. 21,1794–1805 (2009).

23. N. A. Zeliadt, L. J. Mauro, E. V. Wattenberg, Reciprocal regulation of extracellularsignal regulated kinase 1/2 and mitogen activated protein kinase phosphatase-3.Toxicol. Appl. Pharmacol. 232, 408–417 (2008).

24. M. R. Ahmadian, L. Wiesmüller, A. Lautwein, F. R. Bischoff, A. Wittinghofer, Structuraldifferences in the minimal catalytic domains of the GTPase-activating proteinsp120GAP and neurofibromin. J. Biol. Chem. 271, 16409–16415 (1996).

25. G. G. Brownbridge, P. N. Lowe, K. J. Moore, R. H. Skinner, M. R. Webb, Interaction ofGTPase activating proteins (GAPs) with p21ras measured by a novel fluorescenceanisotropy method. Essential role of Arg-903 of GAP in activation of GTP hydrolysis onp21ras. J. Biol. Chem. 268, 10914–10919 (1993).

26. J. F. Eccleston, K. J. Moore, L. Morgan, R. H. Skinner, P. N. Lowe, Kinetics of in-teraction between normal and proline 12 Ras and the GTPase-activating proteins,p120-GAP and neurofibromin. The significance of the intrinsic GTPase rate indetermining the transforming ability of ras. J. Biol. Chem. 268, 27012–27019(1993).

27. K. Rittinger, P. A. Walker, J. F. Eccleston, S. J. Smerdon, S. J. Gamblin, Structure at 1.65Å of RhoA and its GTPase-activating protein in complex with a transition-state ana-logue. Nature 389, 758–762 (1997).

28. S. Lu, H. Jang, R. Nussinov, J. Zhang, The structural basis of oncogenic mutations G12,G13 and Q61 in small GTPase K-Ras4B. Sci. Rep. 6, 21949 (2016).

29. N. Normanno et al., Implications for KRAS status and EGFR-targeted therapies inmetastatic CRC. Nat. Rev. Clin. Oncol. 6, 519–527 (2009).

30. T. E. Stinchcombe, C. J. Der, Are all KRAS mutations created equal? Lancet Oncol. 12,717–718 (2011).

22130 | www.pnas.org/cgi/doi/10.1073/pnas.1908353116 Rabara et al.

Dow

nloa

ded

by g

uest

on

Apr

il 2,

202

1

https://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1908353116/-/DCSupplementalhttps://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1908353116/-/DCSupplementalhttp://swift.cmbi.ru.nl/gv/dssp/https://www.pnas.org/cgi/doi/10.1073/pnas.1908353116

31. L. Gremer, B. Gilsbach, M. R. Ahmadian, A. Wittinghofer, Fluoride complexes ofoncogenic Ras mutants to study the Ras-RasGap interaction. Biol. Chem. 389, 1163–1171(2008).

32. W. J. Jeong et al., Ras stabilization through aberrant activation of Wnt/β-cateninsignaling promotes intestinal tumorigenesis. Sci. Signal. 5, ra30 (2012).

33. S. Grünert, M. Jechlinger, H. Beug, Diverse cellular and molecular mechanisms con-tribute to epithelial plasticity and metastasis. Nat. Rev. Mol. Cell Biol. 4, 657–665(2003).

34. S. Vartanian et al., Identification of mutant K-Ras-dependent phenotypes using apanel of isogenic cell lines. J. Biol. Chem. 288, 2403–2413 (2013).

35. P. D. Adams et al., PHENIX: A comprehensive Python-based system for macromolec-ular structure solution. Acta Crystallogr. D Biol. Crystallogr. 66, 213–221 (2010).

36. A. J. McCoy et al., Phaser crystallographic software. J. Appl. Crystallogr. 40, 658–674 (2007).37. P. Emsley, B. Lohkamp, W. G. Scott, K. Cowtan, Features and development of Coot.

Acta Crystallogr. D Biol. Crystallogr. 66, 486–501 (2010).38. A. Morin et al., Collaboration gets the most out of software. eLife 2, e01456 (2013).

Rabara et al. PNAS | October 29, 2019 | vol. 116 | no. 44 | 22131

APP

LIED

BIOLO

GICAL

SCIENCE

S

Dow

nloa

ded

by g

uest

on

Apr

il 2,

202

1